Qie Niangao
-
Posts
13,445 -
Joined
-
Last visited
Content Type
Forums
Blogs
Knowledge Base
Blog Comments posted by Qie Niangao
-
-
Like Nika, I'm puzzled that the price hike for non-profits wasn't communicated in any form to larger customers in advance of the public announcement. Perhaps that's an idealistic approach -- everybody learns at the same time, a level playing field -- but it's an unusual way to do business. (Pro tip: Reserve good news for surprises; bad news should never come as a surprise.)
It's possible, I suppose, that this is advance warning of some more dire turn ahead. We may all look back and think: "Those lucky non-profits, they got out while the getting was good."
Surely this wasn't an overnight decision. The announcement has very little rationale for the decision, so the reasoning behind it is left to our imagination. (Your World. Your Imagination.®) So, one might imagine:
Perhaps the discounts were always too deep. Is this true? I honestly don't know, but they seemed incredibly generous, to an outsider. Do all suppliers of non-profits typically provide such attractive discounts?
Still, though, if the discounts were just too big, why didn't LL merely scale them back to an industry-standard level (whatever that may be)?
Perhaps the abuse was just too widespread and too difficult to control. It may be hard for "good non-profits" to get their heads around this, but there's a dirty little secret: not all non-profits had clean hands in all their SL dealings. In general, there are 501©3s exploiting some gaping "art", "education", and "research" loopholes in the registration criteria. Add to this the SL-specific practice of having paying tenants of non-profit sims. Even some of those "good" non-profits had such tenants, on the premise that rent payments were just donations to the non-profit. That may sound reasonable, but consider the reductio ad absurdum: for-profit tenants and non-profit sim owners in fact could be the same person. Corporation as "alt" in legal real life.
This kind of nonsense had been going on for a long time, deeply distorting the SL market. Perhaps it was spreading. ("You're paying full price for your sim? Are you crazy? Let me show you how to save a fortune!") Perhaps it was bad enough to jeopardize whatever advantages LL was getting from the discounts.
But this is all speculation. It would be nice if LL could clarify the reasons behind the change.
-
... or: How I Learned to Stop Worrying and Love the Bomb
-
Kwamey I read someplace that the Open Beta testing will only be for a few weeks. Something like 3 or 4 weeks before Mesh Import will be available to everyone on the main grid.
I dearly hope this is not correct. Not because I don't want to see Mesh on the Main Grid, but rather because I don't want Mesh to be another charlie foxtrot of Viewer 2 proportions.
-
Dari, while it's hardly a substitute for proper documentation, the last part of this video demonstrates the most basic possible rigging to the avatar skeleton, done in Blender. This can become much more complex, rigging a mesh to as many as four different bones (if Nyx recalled that limit correctly).
-
Thanks, Nacy. This should keep me busy for a while!
-
Hmm. I'm guessing you mean Domino Marama because I was linked to his website from some other Mesh-related post somewhere, but I haven't found the AV model anywhere there. Undoubtedly it's just hysterical blindness on my part, but if anyone could direct me, I'd appreciate it.
-
It would be best if comments were closed on this ungainly thread. For anyone who may have missed it, Jack has a new blog post announcing the open beta, to start 13 October.
-
Has one of the beta testers yet made available a "stock" rigged human avatar model, for experimentation? There's really no need to wait until the beta is open; a downloadable .dae would give us all something to fiddle with in Blender between now and when the floodgates open.
Speaking of which, I hope there are growth plans for Aditi. Once the open beta starts, I'd expect more than the usual number of users on that grid -- by about three orders of magnitude.
-
Thanks for that information, especially the rendering of Physics Shape metadata. What I was rambling about was an uploaded physics shape, not one created (semi-)automatically at upload. It appears those can differ quite a bit, but there may be constraints... and there probably should be.
One such constraint that I can imagine is that the models' bounding boxes must all have the same center point. (In fact, I believe I read that somewhere, and that the collada model origin is not preserved at import, but replaced by the center of the bounding box.)
If the physics model can differ too dramatically from the mesh appearance, one might devise all sorts of nasty tricks. Imagine, for example, the new landowner's confusion when she can't get anywhere near her property despite there being no visible encroachment. The ability to render Physics Shape might come in very handy in such circumstances.
-
I'd guess that LL plans for the Mesh physical model to be adequate for physical vehicles, but I'd also guess that they're still tweaking the details of doing that efficiently and adjusting the physics "cost" formula accordingly.
A related question, however: Are there constraints on how dramatically a Mesh physics model can differ from the Mesh appearance model? For example, can one "simplify-out" the wheels and undercarriage of a ground-based vehicle Mesh, or indeed make the whole thing effectively phantom? That way, one could link in standard prims with known physics behavior if, worst case, Mesh physics turns out to be inefficient or otherwise problematic. (I'm guessing that the bounding boxes of the physical and appearance models must at least intersect... otherwise this could get "interesting".)
-
Jennifur and Qui and a few others will be HUGE WINNERS in SL in selling some amazing meshes.
Eep! Assuming "Qui" is me, I should clarify that I wasn't in the beta group, and am still only pulling together what I can learn from what beta testers have said and what Lindens have answered at office hours.
So please don't take anything I say about SL Mesh as coming from personal hands-on experience: I don't have any.
I guess, inasmuch as the NDA is gone, LL shouldn't object to my admitting that I was invited to be a Mesh beta tester. The invitation itself was supposed to be kept secret, independent of signing the NDA, and it was the NDA that kept me from participating. (My view of NDAs is about the same as John Dvorak's.)
In case anybody wonders why I'd get invited despite having never sold a sculpty and having quit making them for myself, it was probably because I was a member of the Battery Street Irregulars testing group. (Yeah, this was way back before the axe fell on BSI.) There's no way LL could have known that in a former life I was on a standards committee for CAD/CAM product data exchange.
@Fenrir:
Again, let's go back to sculpties. I remember someone writing a patch for the viewer which allowed you to create sculpties using a raytracing bake system. I believe there was another that allowed direct manipulation of the vertices. Neither were popular or made it into the final viewer code, despite a lot of positive response. Linden Lab didn't seem to be interested in going down that Swiss Army knife road. Those who already had external methods for creating sculpties were using just that -- Blender or Maya or what-have-you -- and they had no need for a built-in basic sculpty creator. It would be a waste of LL's time to put one in SL as only a very small percentage of its population would find it useful.
Right, so this is one model of SL: a marketplace of content between creators (the "very small percentage") and consumers (everybody else). That's a popular model, and growing more popular. Maybe it's the "correct" model for attracting lots of social networking participants, and therefore the correct model for LL.
Judging by remarks here and elsewhere, it's certainly the model held by most beta participants with anything to say on the matter -- and that's not surprising when you think about it.
But that's not my model of SL. To me, it's not the lack of pretty pixels that limits new sign-ups. To me, a new feature such as Mesh is only interesting to the extent that it gives people something to do in Second Life. Something to do with their time, and then their land. Something that makes more compelling the experience, not just the view.
We won't get viewer-based mesh editing because LL developers, feeling sorry for themselves, have rationalized that it's a waste of their time (see Nyx's office hour) because third-party tools are better than what they could build. Fine: they're betting everything on the pretty pixel premise.
Watch. Concurrency will drop slightly as a result of Mesh. The economy will stay more or less flat, with more L$s sloshing toward mesh for a while, but the overall pool won't grow because of this. The only thing that will grow the economy (and concurrency, and LL revenue) is more to do in Second Life, appealing to more people.
Mesh, as planned, is a net-net no-op.
-
Not to mention the standard hand animations, which are buggy as hell. Unfortunately, the best scheme I've heard so far for animating meshes under script control will involve some pretty unintuitive legerdemain in the model itself, and certainly won't look natural.
I still think that the much higher shape detail and especially texture resolution with attached mesh will inevitably replace standard human avatars -- except head and hands, which will still get normal skins from the same people who will sell the skin textures for the attached mesh avatars. And their clothing. Mesh clothing for all but the loosest garments will have to be custom shape-matched with the avatar anyway, whether the avatar's shape was set by slider or in an attached mesh model.
-
As for the mesh beta SIMs, I *think* they are still closed until they roll out the open beta. But, you could ask Nyx Linden to be sure about that. I could be wrong.
Aren't Mesh testers also using a special viewer?
FWIW, I wikified a transcript of Nyx's office hours from yesterday and linked to it from the Mesh article, which is now being very actively reworked by the doc team. I commented on highlights in the Building forum thread. (Jive handles forum threads differently: in "flat" unthreaded view, it splits them into pages, so reloading doesn't take forever and consume a ton of bandwidth. I suspect that's why we keep having problems with expired tokens, eating our posts here.)
(Also, apropos emails from here: unsubscribe only works until you post another comment, as long as you have the default "on" setting for Your Stuff / Preferences / Email Notification Preferences / "Automatically receive notifications for blog posts I comment on." Hey, it wasn't my idea.)
-
@Toysoldier (and others): There's a thread in the Building forum which may be of interest. I've gotten responses to technical questions there that just got buried here, so if you need details you can't find in the various Mesh blogs or elsewhere, that's another source.
From comments there, it appears that a very great deal is still TBD, including what I should think would be pretty fundamental aspects of the implementation. In that light, it's likely impossible at this point to really know such pure "economic policy" things as scale limits, upload fees, prim-equivalence constants, account restrictions, etc.
On IP infringement: The state-of-the-art in automated detection of copied content is a lot higher than I appreciated. Google has implemented some truly amazing ability to detect YouTube copies of degraded snippets of video streams -- even to the point of routinely finding 15-second clips of week-old network TV broadcasts, partially obscured and at an oblique viewing angle, shown in only part of the frame. (Think about that for a moment.)
So, I'm now thinking: Bring it on. Let the sleazeballs copy everything on the web. Let LL's lawyers finally face the fact that they're in exactly the same business as Google is with YouTube, and have exactly the same legal obligations. And they can use available technology to protect SL content creators just like they'll have to protect external content.
(Yes, there are economic implications of fully legal importing of content. That's just not what I'm talking about here.)
About who's more 1337 than whom: First of all, just because you don't recognize sophisticated prim builds, or don't know how to make them yourself, doesn't mean they don't exist. And if you think it's easier to wrangle prim geometry after the fashion of Siefert Surface than it is to unwrap a UVW map from a modelling program, you're deluded.
Moreover, if you think anything will prevent amateurish Mesh content from appearing in Second Life, you better pray you're wrong or SL is over and it's time to turn out the lights.
-
Yeah, maybe a linkset exporter as a stopgap, or even a scripted in-world "mesh builder" that exports models for manipulation in Blender. Okay, Reed is right and I was too much "a Negative Nancy" about that, but I've used in-world sculpty editors, and animation tools, and while useful for specialized applications, they truly do suck. Anyway, for my purposes, it could only be a stopgap until some virtual world comes along that doesn't cripple its building environment.
I actually like Blender, and would choose it over any of the other external 3D modelling programs I've tried. Yeah, the UI is quirky, but the quirks are not without rationale. The problem is that Blender is not SL. The more my workflow depends on Blender, the less relevant and compelling SL becomes for me. I went through all that with sculpties.
I recognize that I'd feel differently if my objective were to generate built stuff to sell, with SL as primarily a marketplace for those builds. Indeed, that sentiment was likely heavily represented in the mesh closed beta.
It's extremely frustrating to keep seeing 2D graphics (and even sound ) raised as examples of things that can't be done in SL that are intended to be somehow analogous to editing a 3D model. Did anybody come to SL to edit 2D graphics in-world? Is it so difficult to imagine there are folks like me who came to SL for an immersive social 3D modelling experience?
Granted, I have to be willing to accept that the right business decision for LL may be to just cut loose folks like me, and favor the market of "builds" over that of "building." Probably the Facebook masses and many of the merchants who cater to them are in that former market, at least for now. So folks like me may have to go elsewhere, until creating 3D content becomes as much what's expected of social environments as typing status updates is today.
-
I understand why poeple would *want* a mesh editor in world. I also would ask: "you willing to wait?".
Oh, I'm willing to wait.
I'm not sure the competition is.
Usually I'm the last person to raise the spectre of competition, but if I have a choice between spending time building directly in a virtual world that supports it, or building in Blender for later upload to SL, where do you think I'm going to spend my time?
Even if the tools are extremely basic and limited, even if the platform itself is rough and more than a little makeshift, I'll build inside that social environment every time I possibly can.
I get it that there are people who, in contrast, get themselves a standalone OpenSim to build in isolation. That's not my idea of the future of virtual worlds -- and I really think LL had better hope that's not the future of virtual worlds.
As to mesh editing appearing first in (possibly specialized) third party viewers because LL development is too backlogged and process-constrained: Works for me. I suggested it in an earlier post. LL would still need to expose messaging for mesh manipulation between the sim and viewer.
About a more extensible, plug-in architecture for the viewer: Great. We've been wanting that for years, and its lack is one of the great disappointments of Viewer 2, considering the vast resources that were poured into that project: Opportunity cost, in spades. Hence, I wouldn't interpose an extensible viewer project on the critical path for anything I wasn't trying to kill. And in any case, even if the viewer were perfectly extensible, again, LL would need to expose messaging for mesh manipulation between the sim and the viewer.
I'd feel a lot better (and STFU) about all this if viewer-based mesh editing were on a development roadmap somewhere ahead of all the tweaks and low-impact features that folks will dream up for mesh. I saw what happened with sculpties -- thousands of man-hours sunk into "nice to have" features and performance tweaks, now as near to irrelevant as makes no difference. Unless mesh editing gets a stake in the ground up-front, the opportunity will be frittered away, again.
-
“Oh no! But there aren't any in world tools to make meshes like there are for sculpties!”
Yes there are. A couple. They're already made.
And they will suck just as badly as the ones for sculpties.
Most of the code needed for Your Favorite Mesh Software is already in SL. The SL mesh viewer has, of necessity, everything it needs to display meshes (including texturing, lighting, etc), and to translate, rotate, and manipulate their scale in three dimensions. Sure, that's not the modeling functions, but it's not as if adding a mesh editor to a viewer would be adding one to a spreadsheet, or starting from scratch.
Some people just don't get why others of us want in-world mesh editors. It is simply the difference between those who like buildings in SL and those who like to build in SL.
-
Something came up at Kelly's office hour that has me confused, based on what I thought I understood from discussions at Nyx's office hour last week and other sources.
I now gather that an uploaded mesh in SL is a directly rezzable thing, not something that's applied to a rezzed object as one might apply a sculptmap. That raises a couple of questions, if anybody knows:
- Is there any way to change the mesh used in an object instance, the way we can change the sculptmap in a sculpty? (That's how I'd assumed mesh "animation" might be possible.)
- If one were to make a hundred copies of a mesh object, would that be a hundred separate mesh definitions, each separately downloaded to an observing viewer? Or is there some shared mesh geometry asset that one can access (and use, for example, to address #1 above)?
-
The secrecy was important because as mesh as developed in SL the methods of importing it and how it functions were changing. There was no reason to release information that was very likely to change and in many cases did and to some extent is still being tweaked before the public beta. The past three months have seen changes to how mesh is imported and how many prims its going to cost (still being tweaked), how its physics will be handled and the addition of rigging for attachments/avatars.
(My emphasis.) That is to say, for the purposes of public release of information, we're still in almost exactly the same place we were three months ago, or even earlier: We still don't know how prim equivalence will really be calculated, except that it's the max of some TBD constant applied to file size and some TBD formula for physics cost.
I need to grep my old chatlogs for the Linden quote about the (re-)imposition of the NDA being all about Mesh's impact on the SL economy.
I seriously doubt any of these new avatars will replace the existing SL avatar as a standard...there's just way too much content that would be broken.
No. Just no. I mean, really: No fashionista is going to wear a crappy system avatar (soon to be labelled the "n00b avatar") and stand next to another fashionista wearing ten times as many vertices and a hundred times the skin resolution. Not a chance in hell. The existing skin and clothing assets are simply obsolete.
The sooner skin and clothing creators absorb that fact, the less likely they'll be bulldozed into oblivion by Mesh. There's a huge opportunity here and somebody is going to take advantage of it.
-
Have you used in-world sculpty editors? It's not a terrifically satisfying experience. To say nothing of the Rube Goldberg data path to produce an actual sculpty with this approach.
And really, Ann, you of all people must realize that the days of 1.x-based viewers are numbered. Henri's Herculean efforts notwithstanding, the "big boys" are already re-basing on 2.x Snowstorm, which will be a great deal easier than backporting Mesh (and all the rest of Viewer 2 that hasn't been backported yet). It remains to be seen how many UI fixes LL will adopt in Snowstorm and how much will rely on TPV developer initiative outside Snowstorm.
(Tangent: I'm watching what's happening with detachable sidebar panels right now; the current Snowstorm implementation completely misses the point, and at least one proper patch has been submitted, supporting such configurations as multiple simultaneous profiles and group info panels. If Oz&Co don't pick that up, and never tell us why not, then LL viewers have no future.)
-
Can any linden or tester comment on how lip-sync and/or emotes will work with avatar meshes?
I'm neither of those, but it's been discussed, including at Nyx's office hour. Facial expressions for emotes and lip-sync are not supported in mesh. (Same with hand position morphs.) Across the street, we've speculated that this will mean human avatar heads will survive the transition to attached-mesh avatars. On the plus side, the system head mesh is better than the rest of the avatar and has the advantage of a dedicated full-res texture for its standalone UV map.
Of course, for non-human avatars, that's not helpful. I don't know of any information about when or if rigged meshes will get shape keys for morphing.
-
I've heard the "but SL must have a built-in editor for this new feature!" argument before, with sculpties. I'm not convinced, and LL rightly ignored the requests. And what happened? People created in-world sculpty building tools.
Yes. And even though the in-world sculpty builders were clumsy, slow, and in every way inferior to external programs, people used them. Now, why might that be?
Since the introduction of sculpties, the papers for a divorce between SL content and in-world editing were already on court record.
Since forever, for content other than the built model. But yes: sculpties started us down the track where it was no longer possible to build while immersed in the environment for which one was building.
Anyway, like with sculpties, I am sure very enterprising and creative residents will come up with in-world Mesh editing programs.
Unless LL exposes a viewer messaging format for direct manipulation of temporary Mesh assets, those in-world programs will suck in all the same ways as do in-world sculpty editors -- and in-world animation tools.
There's higher initial up-front network cost for downloading a Mesh, but once you have the 3d mesh and its associated animations in your cache, animating it costs no more network bandwidth -- Unlike animating sculpties or prims in SL now.
Umm. Huh? What "animations" are these? As far as I know, animating a Mesh object will be precisely like animating a sculpty, except that it won't be possible to use a transparent texture to precache alternate meshes, as we can with sculptmaps. (Or... maybe this refers to avatar animations deforming rigged Meshes... which is cool and all, but doesn't seem to be what this is about.)
Anyway, look, I'm very well aware that I'm not going to get my in-world, viewer-based Mesh editor any time soon. I knew that wasn't what the closed beta was testing, but when there was a brief "NDA leak" and then absolutely nothing for months, during which time Philip coined "Fast, Easy, Fun", I had a glimmer of hope that they were taking SL back to its roots as an immersive 3D creation environment.
It's now quite clear that nothing of the sort was happening. And that brings me back to my much earlier post, the fourth in this thread: Just what the hell have they been doing for the past three months?
Back around the time of the NDA leak, some Linden mentioned that all the secrecy was necessary because Mesh would have a very profound effect on the SL economy. That stuck in the back of my mind: Yeah, it's a big effect on the economy, but how does it help to keep the technical details secret all that time? Were they going to use that time to develop a differentiator from competitors? Something like an in-world Mesh editor, perhaps? Clearly not.
My best guess at this point is that the secrecy was to make sure people wouldn't stop buying skins and texture-based clothing for their soon-to-be-obsolete system avatars (and sculpties, and prim builds). It would be bad if the economy had undergone a six month hiatus while everybody awaited the vastly superior content. That's just not the reason for which I'd hoped.
-
Sure the old guard are pissed off, most of them thought they could keep selling their 2005 model prim based stuff for another 20 years, sure it's some of the best stuff made with limitations they had, but they don't want SL to advance, they want to kep the limitations so they can stay on top with no additional work, if some had their way we'd be overtaken by Active Worlds and would be lucky to have better than 8 bit color...........
People keep saying that prim builders are foot-dragging, and ascribing to them all sorts of nefarious motivations for their anachronism. I just don't see them saying that in this thread, so I'm not sure where people are getting it. (I refuse to look in the Commerce forums cesspool, so maybe there?)
I absolutely do see folks wanting to be able to build in-world using current, viable model types, and now Mesh is a requirement for that. Others disagree, but I really believe that in-world Mesh editing is tablestakes for SL's survival, and sooner rather than later.
That's not to say there aren't other risks posed by Mesh; I've mumbled about some in this thread, other posters have addressed different risks. There's valid reason to be concerned about some of these, especially right now.
For example, widespread illegal import of commercial models will push LL further along the knife-edge it's been straddling between non-compliance with DMCA requirements on one side, and on the other, DMCA as the most effective anti-creator griefing tool anyone could imagine. We just have to hope that LL's legal department is ready.
Also, the SL economy is pretty fragile right now, especially LL's land-based revenue (which is the vast majority of its income). Maybe Mesh will entice more "residential" folks to really need a plot of land to rezz an ever-so-pretty Mesh house and fill it with ever-so-detailed Mesh trinkets. And some in-world stores will sprout up, carrying Mesh products (one hopes the need to see Mesh products in-world to appreciate them will be enough to motivate Merchants to pay tier for "brick and mortar" and not just follow the growing herd to Marketplace-only retail). But will these new in-world land uses outweigh the inevitable retreat from acres of floorspace where old-school skins and clothing are sold, combined with the builder workshops emptied as interest wanes for in-world building? Changing demand for land is a real risk for LL.
Another thing I've not seen mentioned is that Mesh makes SL largely build-compatible with Blue Mars (among others, now and in future). This is an opportunity some creators will appreciate, but it represents a vulnerability for LL, lowering the barrier for exodus. If, for example, all of Caledon were built with Mesh, I wonder how much would remain in Second Life, and for how long.
-
The upload cost doesn't really matter. Anything that's going to require multiple uploads to get right can be done on the Aditi grid for free, and anything not worth the L$ price to upload the final version really shouldn't be cluttering up the asset server.
There are very good reasons that in-world editing is essential to SL's future, but cost of upload isn't really one of them.
(Just for the tinfoil millinery crowd: "Revenue" from uploading mesh is very unlikely to be much of a consideration for LL, either, because it won't be that much, and it will be in L$ "funny money" only monetized as a "sink" to be filled by Supply Linden sales, maybe, someday.)
You also mentioned animations, which may play out very differently. I'm hearing vague rumblings that some variant of the "expressive puppeteering" bit of the old Physical Avatar project is slated for a comeback. If so, it will change everything about animation, and make interacting with SL avatars much more compelling than we could ever hope to achieve with a shapelier mesh.
Facebook
Instagram
Twitter
YouTube
Flickr
Two Important Updates on 2011 Land Pricing
in Land
A group blog by Secondlife in General
Posted
Dean: "So buy a server and talk to I.T. about support. On your three-year replacement cycle that's a lot cheaper. Get one of our in-house grants to attend a training class on running this OpenSim thing."
Hmmm. Considering what I've heard about the accounting practices in education and non-profits, I can see that actually happening. But it certainly wouldn't advantage the institution as a whole.
It's very expensive to introduce a new and idiosyncratic application, running on standalone hardware, with heavy network demands, unspecified security needs, frequent updates, and "support" by self-serve source code inspection. Way, way, way more expensive than the undiscounted LL fees.
But if a department can hide its real costs in some other (IT) organization's budget, it very well might, assuming that IT organization is dumb enough to take it on.
So, what are the options, really? I can think of four:
It's probably obvious that I find the fourth option most interesting, inasmuch as there's at least the hope of some silver lining there. There are surely other alternatives. This would seem a good time to constructively explore creative approaches to dealing with the new situation.